How Do Pollsters Accurately Measure Latino Opinion?

How do pollsters accurately gauge Latino voter interest? We wrote last week that there was a question as to whether polling based on Spanish surname was the best methodology, or whether there were other factors at play that made a big difference.

That’s a good question, a couple of top Latino pollsters got together on Twitter this week to hash it out over the #pollingwars. This included Matt Barreto and André Pineda. Carlos Odio over at the Daily Grito spelled it out, here’s an excerpt:

Barreto says you need Latino household lists culled from various sources, and you need interviewers calling them who are ready to survey voters in Spanish as well as in English.

Pineda seems to agree. He also thinks you need to call a wider, random sampling of voters who aren’t identified as Hispanic in the voter file and ask whether they self-identify as Latino. Whether we’re missing 3% of the Latino electorate or 20%, accuracy demands that we “minimize shortcuts,” he says.

What we haven’t mentioned are considerations of cost. Random-digit-dialing of voters in a precinct actually sounds like a better way to get closer to a more representative sample of Latinos. But it’s expensive. Really expensive. As Barreto mentioned in one of his memos, the pollster responds to the preferences of his or her clients. You respond to what the client values, what they’re willing to pay for.

There’s more specific analysis on the original post, check it out and let us know what you think!

Follow Sara Inés Calderón on Twitter @SaraChicaD